Cervical cancer is one of the top four common predictors of cervical cancer related mortality among women worldwide, with an estimated 570,000 cases and 311,000 deaths in 2018[1]. According to global cancer related death statistics, cervical cancer death was 18 times higher in low and middle income countries than in developed countries, which accounted for 85% of cervical cancer mortality [2]. Cervical cancer is caused by oncogenic subtypes of the human papillomavirus [3]. Cervical cancer accounts for 7.5% of all female cancer deaths worldwide, peaking between the ages of 35 and 65, and not only kills young women but also destroys families with young children [4, 5].
As a result of population ageing and growth, as well as an increase in the prevalence of risk factors associated with economic change, such as smoking, obesity, physical inactivity, and reproductive behaviors, cancer is becoming more prevalent in Africa[6]. It is the leading cause of cancer-related death in women in Eastern, Western, Middle, and Southern Africa [7]. Cervical cancer treatment in Africa is hampered by a lack of diagnostic and treatment facilities, a lack of healthcare infrastructure, and insufficient pathology services [8]. Cervical cancer is expected to kill over 443,000 women worldwide by 2030, the vast majority of whom will be in Sub-Saharan Africa [9].
Each year, approximately 6,300 new cases of cervical cancer are diagnosed in Ethiopia cancer treatment centers, and approximately 4,884 women die from the disease [10]. According to data from TikurAnbesa Specialized Hospital's (TASH) radiotherapy centre, cervical cancer is the second most common type of female cancer among patients attending the oncology centre [11]. According to various sources, treatment options include surgery, chemotherapy, radiotherapy, or a combination of the three [12–14].
The main goal of this study was to assess the pooled size of cervical cancer mortality and its predictors in Ethiopia.
The study's findings will also be used to raise public awareness about predictors of death among cervical cancer patients. It also allows us to share our findings with the Ethiopian Ministry of Health in order to help policymakers raise public awareness of cervical cancer-related death, paves preventive and curative aproaches for early detection and proper treatment.
Setting
Ethiopia is a country in East horn of Africa, divided into ten regions: Tigray, Afar, Amhara, Oromia, Somali, Benishangul-Gumuz, the Southern Nations Nationalities and People Region (SNNPR), Gambella, Harari, as well as two administrative states (Addis Ababa city administration and Dire Dawa city administration)[15]. In the 1960s, Ethiopia had a population of twenty million people. Ethiopia has a population of over 120 million people, making it Africa's second most populous country next to Nigeria [16].
Sources of data and ways of search
Common Public databases like Science Direct, Embase, the Cochrane Library, and PubMed were thoroughly searched for studies that explained the predictors and magnitude of cervical cancer patient mortality in Ethiopia. Endnote software citation package was used to cite and download articles. To find studies, the following keywords were used: "survival of cervical cancer patients," "magnitude of cervical cancer related death," "epidemiology of cervical cancer mortality," "Ethiopia," "determinants of cervical cancer death," and "factors related to cervical cancer survival." The search terms were used individually as well as in conjunction with Boolean operators such as "OR" and "AND." Citations discovered using our search terms were imported into EndNote-X6 software, and duplicate articles were removed.
Inclusion and exclusion criteria
The review attempted to include published studies with epidemiological data on cervical cancer mortality and predictors conducted in Ethiopia between Jun 2003 and 30 March 2023. To assess the paper's quality, the full texts of selected articles were retrieved and carefully appraised. Other articles published in English online from university repositories were also considered. Newcastle-Ottawa Scale assessment of Critical Appraisal Checklist for Cohort Studies was used to determined quality status of the article and included articles were accepted at the level of five and above score out of ten overall score[17].The screening was based on methodological issues, incomplete data, or full text that was not accessible from the analysis. [18].Then, described by Preferred Reporting Items for Systematic Review and Meta-analysis protocol (PRISMA-P) Fig. 1
Information extraction and quality assessment of included studies
Three researchers extracted data independently after selecting articles of interest for a complete text review using a Microsoft Excel spreadsheet. The name of the author, publication year, area or region, study design, age range, sample size, response rate, number of study subjects with the outcome of interest, prevalence rates, and predictors significantly related to cervical cancer-related Mortality were all included in a Microsoft Excel spreadsheet.
Characteristics of included studies
This meta-analysis included 14 studies that reported on the magnitude and predictors of mortality among cervical cancer patients in Ethiopia. Totally, 9,260 cervical cancer patients were included in the study. Articles published between January 2013 and March 30, 2023 was taken into account. This study looked at the most common predictors of mortality in cervical cancer patients. Those predictors were, late diagnosis [19–22], radiation therapy only[11, 23, 24], surgery with adjuvant therapy [25–27] and anemic status of the patient[28–30] were selected based on their frequency identified as the predictors by different studies (Table 1).
Table 1
Summary of studies involved in meta-analysis
Author Name | Publication year | Study Design | Total Sample | Total Events (%) | Response rate (%) | Quality score |
Aguade et al | 2023 | Cohort | 322 | 48 | 98.0 | 9 |
Argefa et al | 2022 | cohort | 175 | 33 | 91.0 | 7 |
Begoihn et al | 2019 | cohort | 1495 | 52 | 95.0 | 8 |
Deressa et al | 2021 | cohort | 242 | 57.9 | 89.0 | 8 |
Fikreaddis T et al | 2018 | cohort | 252 | 5.6 | 98.5 | 8 |
Gashu et al | 2023 | cohort | 322 | 36.60 | 92.0 | 8 |
Gizaw et al | 2017 | cohort | 1655 | 52 | 89.0 | 9 |
Gurmu M et al | 2018 | Cohort | 907 | 38.5 | 90.0 | 8 |
Kantelhardt et al | 2014 | cohort | 1,059 | 23.5 | 98.0 | 9 |
Mebratie et al | 2022 | cohort | 422 | 27.66 | 89.0 | 7 |
Mölle G et al | 2016 | cohort | 1009 | 49 | 92.0 | 8 |
Olyad M et al | 2021 | cohort | 418 | 9.9 | 91 | 8 |
Seifu et al | 2022 | Cohort | 348 | 31 | 99.4 | 7 |
Wassie et al | 2019 | cohort | 634 | 38.62 | 98.0 | 8 |
Statistical analysis and Model selection
The heterogeneity inverse variance (I2) test was used to assess study non-uniformity, and Cochran's Q test (P-value less than 0.1) revealed statistically significant results. [31]. The I2 value ranges from 0 to 100%. I2 75% indicates that there is significant heterogeneity across studies, with a P-value of 0.05 used to declare significant heterogeneity [32]. The pooled size magnitude of cervical cancer related mortality was estimated using a random effect model because I2 = 78.50 at p = 0.00, indicating the presence of heterogeneity between studies. Furthermore, we used a random model for subgroup analysis by identified predictors to deal with the significance of their significance in cervical cancer-related mortality at I2 = 96.4%, indicating high heterogeneity across studies [33].
Publication bias Assessment
A funnelplot graph was used visually to assess publication bias. The symmetry of the funnel plot revealed that there was no potential publication bias [34] (Fig. 2). Egger's and Begg's tests were also run at the 5% significant level; the distribution of each study, as well as a P-value greater or less than 0.05, were used to determine the presence or absence of publication bias [31].(Table 2)
Table 2
Egger’s test for small study effect
Number of studies = 14 Root MSE = 10.49 |
Stf_Eff | Coefficient | Std_error | t-value | p>|t| | 95% confidence Interval |
Slope | 31.33 | 15.75 | 1.99 | 0.070 | -2.99 65.65 |
Bias | 2.06 | 9.09 | 0.23 | 0.83 | -17.76 21.88 |
Test of H0:no small study effects P = 0.825 |
The Egger's test output also indicates that there is no publication bias at (P = 0.825). However, the result of Begg's test was notable in the presence of minimal publication bias (P = 0.047).Trim and fill analysis was thus performed following pooled effect analysis (Fig. 2).